A phase I/II trial of vandetanib for patients with recurrent malignant glioma

Abstract

Vandetanib is a once-daily multitargeted tyrosine kinase inhibitor of vascular endothelial growth factor receptor–2, epidermal growth factor receptor, and the rearranged-during-transfection oncogene. A phase I trial was conducted to describe the pharmacokinetics of vandetanib in patients with recurrent glioma on enzyme-inducing anti-epileptic drugs (EIAEDs) and to identify the maximum tolerated dose (MTD) in this population. A phase II trial evaluated the efficacy of vandetanib in patients with recurrent malignant glioma not on EIAEDs as measured by 6-month progression-free survival (PFS6). In the phase I trial, 15 patients were treated with vandetanib at doses of 300, 400, and 500 mg/day, in a standard dose-escalation design. The MTD in patients on EIAEDs was 400 mg/day, and steady-state levels were similar to those measured in patients not on EIAEDs. Dose-limiting toxicities were prolonged QTc and thromboembolism. Thirty-two patients with recurrent glioblastoma multiforme (GBM) and 32 patients with recurrent anaplastic gliomas (AGs) were treated in the phase II trial, at a dosage of 300 mg/day on 28-day cycles. Six patients (4 GBM, 2 AG) had radiographic response. PFS6 was 6.5% in the GBM arm and 7.0% in the AG arm. Median overall survival was 6.3 months in the GBM arm and 7.6 months in the AG arm. Seizures were an unexpected toxicity of therapy. Vandetanib did not have significant activity in unselected patients with recurrent malignant glioma.

Patients with malignant gliomas continue to have a poor prognosis despite standard therapy with maximal surgical resection, radiation, and temozolomide. Patients with glioblastoma multiforme (GBM) treated with chemoradiation have a median survival of 15 months, with those having a methylated methylguanine methyltransferase promoter benefiting more from temozolomide.^1,2 Progression of disease is inevitable, and prognosis at the time of progression is poor, with a median overall survival (OS) of 6 months in GBM and 11 months in anaplastic glioma (AG).³ Clearly, more effective therapies are needed. Toward this end, anti-angiogenic and molecularly targeted therapies are increasingly being explored as potential advances in glioma management.

Approximately 40%–50% of GBMs exhibit dysregulation of epidermal growth factor receptor (EGFR), which has been associated with poor prognosis and response to treatment.^4–6 In particular, EGFR variant III (deletion of exons 2–7) leads to ligand independent activity that drives GBM growth.⁷ EGFR inhibition with gefitinib and erlotinib monotherapy has, however, demonstrated only marginal benefit in recurrent malignant gliomas—objective response rates range from 0% to 25%, with minimal effect on long-term outcomes.^8–10

Anti-angiogenic strategies are a promising approach for treatment of malignant gliomas secondary to the highly vascular nature of these tumors, and preclinical data demonstrate the dependence of glioma growth on generation of tumor-associated blood vessels.^11,12 Glioblastoma cells express high levels of vascular endothelial growth factor (VEGF) in situ, and inhibition of VEGF signaling impedes growth of glioma xenografts in immunodeficient mice.¹³ Early clinical experience with bevacizumab, a monoclonal anti-VEGF antibody, has demonstrated clear activity, with radiographic responses that had not been routinely observed in these tumors. Most GBM patients, however, do not achieve long-term disease control with bevacizumab therapy, with 6-month progression-free survival (PFS6) rates ranging 29%–36%.^14,15

In acknowledgment of the complexity of oncogenic pathways and their regulation, many investigations in glioma therapy are focusing on regimens that have multiple complementary targets. Vandetanib (ZD6474, Caprelsa, AstraZeneca) is a combination tyrosine kinase inhibitor of EGFR and VEGF receptor (VEGFR)–2 (kinase insert domain receptor [KDR]). Simultaneously blocking phosphatidylinositol-3 kinase signal transduction via EGFR inhibition and blocking angiogenesis via KDR axis inhibition may lead to improved outcomes for patients. Preclinical studies have demonstrated antitumor effects of vandetanib in rat and mice glioma xenografts.^16,17 Vandetanib has also been tested in several solid and hematologic malignancies and recently received FDA approval for metastatic medullary thyroid cancer.¹⁸ Prior phase I studies in patients with solid tumors established a maximum tolerated dose (MTD) of 300 mg.^19,20 These studies excluded patients on enzyme-inducing drugs, since vandetanib is metabolized in part by CYP3A4.²¹ Approximately 50% of patients with WHO grade III tumors and 25% of patients with grade IV tumors have seizures, and some are treated with enzyme-inducing anti-epileptic drugs (EIAEDs),²² which can be potent inducers of CYP3A4. We therefore conducted a phase I study to determine the pharmacokinetics (PK) and MTD in patients on EIAEDs, as well as a separate phase II study to assess the clinical antitumor activity of vandetanib in patients with recurrent malignant gliomas not on EIAEDs.

Materials and Methods

Eligibility Criteria

Phase I

Eligibility criteria included age ≥18 and histologically confirmed GBM, gliosarcoma, anaplastic astrocytoma, anaplastic oligodendroglioma, anaplastic mixed oligoastrocytoma, malignant glioma/astrocytoma not otherwise specified, progressive low-grade gliomas, or infiltrative brainstem gliomas, diagnosed radiographically. There were no restrictions regarding the number of prior relapses. All patients had progression after radiation therapy that was completed at least 4 weeks prior to study enrollment. Other criteria for study entry included Karnofsky Performance Status ≥60, normal echocardiogram, and normal organ function as measured by white blood cell count ≥3000/μL, absolute neutrophil count ≥1500/mm³, platelet count ≥100 000/mm³, hemoglobin ≥10 g/dL, serum glutamic oxaloacetic transaminase and bilirubin ≤2 times upper limit of normal, and creatinine ≤1.5 mg/dL or creatinine clearance ≥60 cc/min. Subjects must have recovered from toxic effects of prior therapy and not received vincristine within 2 weeks, nitrosureas or radiation therapy within 6 weeks, procarbazine within 3 weeks, other prior cytotoxic therapy or surgery within 4 weeks, any noncytotoxic investigational agent within 2 weeks, or other noncytotoxic agents within 1 week of study entry. A stable daily dose of corticosteroids for ≥5 days was required before obtaining the baseline MRI scan. Exclusion criteria for participation were other concurrent tumor therapy, active malignancy, active infection, pregnancy, breast feeding, QTc >460 ms, poorly controlled hypertension (systolic blood pressure > 160 or diastolic blood pressure > 100), active gastrointestinal disease with diarrhea, severe or uncontrolled systemic disease, and clinically significant dysrhythmia or cardiac disease. Patients may have received prior therapy with VEGF-directed agents. All patients in the phase I study were on EIAEDs. Patients were eligible after signing informed consent for participation in this National Cancer Institute Institutional Review Board–approved trial (NCT 00272350).

Phase II

Eligibility criteria for phase II were similar to those of the phase I trial, except that patients with progressive low-grade gliomas or infiltrative brainstem gliomas were not eligible. EIAEDs were not allowed in the phase II study.

Treatment Plan and Endpoints

Phase I

The primary endpoints for the phase I study were to define the MTD and dose-limiting toxicities (DLTs) and establish the PK profile of vandetanib in patients on EIAEDs. Patients received vandetanib at 300, 400, or 500 mg/day on a 28-day cycle. Dose escalations were planned in groups of 3 patients if no patients in the initial group had DLTs. Three additional patients were added in any dose level where 1 of 3 initial patients had a DLT, defined as any nonhematologic grade ≥3 or hematologic grade 4 toxicity (excluding hypertension) observed within the first cycle of therapy as based on the National Cancer Institute Common Toxicity Criteria for Adverse Events (CTCAE) version 3.0. MTD was defined as the dose at which fewer than one-third of patients experienced a DLT. Nonevaluable patients, such as those removed from the study in the first cycle for reasons other than toxicity, were replaced. Dose reduction was not allowed in the first cycle of the phase I trial.

Steady-state vandetanib levels in patients taking EIAEDs were compared with historical data from patients not taking EIAEDs. Heparinized blood samples (7 mL) were obtained prior to the morning dose on day 28, and at 1 h, 2 h, 4 h, 6 h, 8 h, and 24 h after the morning dose on day 28. The blood samples were immediately centrifuged. Plasma was removed and stored at −20°C until analyzed. Plasma concentrations of vandetanib at the various time points were determined by high-performance liquid chromatographic assay as previously reported.²³ The area under the curve (AUC) and the concentration at steady state were derived from the concentration-time curve.

Phase II

The primary endpoint of the phase II study was PFS6. Vandetanib was self-administered at a dosage of 300 mg orally once daily on a continuous 4-week treatment cycle. Dose modifications were allowed for grade ≥3 nonhematologic and grade 4 hematologic drug-related toxicities. History and physical and neurological exams were performed at the end of each cycle and at the middle of the first cycle, whereas hematology and blood chemistries were performed every 2 weeks. A 12-lead electrocardiogram was performed weekly for the first 8 weeks and then every 4 weeks thereafter. Toxicity was assessed using CTCAE v3.0.

Response to therapy was evaluated with perfusion brain MRI at the end of each cycle. Tumor response was assessed with modified Macdonald criteria using the largest cross-sectional diameters of measurable lesions and evaluations of nonmeasurable but evaluable disease.²⁴ Determinations of complete or partial response required a stable or decreasing dose of steroids. Since response evaluations included assessment of both contrast-enhancing disease and T2/fluid attenuated inversion recovery nonenhancing disease, they were comparable to the current Response Assessment in Neuro-Oncology criteria.²⁵ Pseudoprogression was an ill-defined entity at the time of study inception, and 2 patients were enrolled within 12 weeks of completion of chemoradiotherapy. One of those patients had histological confirmation of viable tumor prior to study entry. All responses were independently evaluated by 2 reviewers (T.K. and J.S.), where measurable disease superseded evaluable disease criteria for response, but progression was determined by earlier date of either method. The designation of stable disease required a minimum of 8 weeks’ duration.

PFS was estimated by Kaplan–Meier methodology using the earliest of date of radiographic progression or date off study for clinical decline. Unless the date of progression after removal from the study was known, patients removed from the study for toxicity were censored at the later of either the last evaluation on study or the start date of alternate therapy without disease progression. All patients were evaluable for survival; only those with a follow-up MRI scan were evaluable for radiographic response. OS was calculated from study registration date until date of death, or, if date of death was not known, patients were censored at the last date known to be alive.

At the time of the study's inception, a database of 225 GBM patients enrolled in 8 previous phase II trials of ineffective therapy was used as a historical benchmark for PFS.³ This trial was designed to distinguish between 15% and 35% PFS6 in the GBM arm, with a planned accrual of 32 patients. Failure of at least 8 patients to progress by 6 months would have given a 0.92 probability of concluding that the agent was effective if PFS6 was 35% and a 0.10 probability of concluding that the agent was effective if PFS6 was only 15%. The trial was designed to distinguish between 30% and 55% PFS6 in the AG arm, with a planned accrual of 32 patients. Failure of at least 13 patients to progress by 6 months would have given a 0.89 probability of concluding that the agent was effective if PFS6 was 50% and a 0.13 probability of concluding that the agent was effective if PFS6 was only 30%.

As a secondary analysis, we explored whether several common toxicities associated with vandetanib within the context of this study (rash, hypertension, and seizure) could be used as biomarkers to predict response to treatment. The correlation between clinical variables and OS and PFS was assessed by the Cox proportional hazards model. The correlation between clinical variables and response was assessed by Fisher's exact test for categorical covariates and by logistic regression for continuous covariates.

Results

Phase I

Patient characteristics

Fifteen patients were accrued to the phase I study between January 2006 and August 2007. Three patients in the phase I study were registered but were not evaluable because they did not complete 4 weeks of therapy for reasons other than DLT and were therefore replaced.

Toxicity and MTD

There were no DLTs on the 300-mg or 400-mg dose. There were DLTs in 2 of 6 patients treated on the 500-mg dose; 1 was a grade 3 prolonged QTc, and 1 was a grade 4 thromboembolic event (Table 1). MTD per day was 400 mg. Patients received a median of 2 cycles of therapy (range, 1–12). Ten patients discontinued treatment due to tumor progression and 5 due to toxicity. In addition to the DLTs in the first cycle, 3 patients discontinued treatment for toxicity in subsequent cycles, including CNS hemorrhage, rash, and hypertension.

Table 1.

DLTs in the phase I study

Dose Level (per day)	Patients	DLTs	Replaced
300 mg	3	None	0
400 mg	6	Grade 3 rash	1
500 mg	6	Grade 3 prolonged QTc	2
		Grade 4 DVT/PE

Abbreviations: DVT, deep vein thrombosis; PE, pulmonary embolism.

Pharmacokinetics

PK sampling was performed in 4 patients on the 300-mg daily dose and 2 patients each on the 400-mg and 500-mg daily doses in the phase 1 study. The PK values for vandetanib, including steady-state concentrations, AUC, and dose-normalized AUC, are summarized in Table 2. Figure 1 shows a trend for the steady-state exposure of vandetanib to increase with the dose. There was no significant difference in the dose-normalized AUC data for patients on EIAEDs compared with normalized AUC historical data in patients with systemic cancer treated with 300 mg a day of vandetanib (Fig. 2).

Table 2.

Vandetanib concentration at steady state (C_ss, max) and area under the curve at steady state (AUC_ss) in patients on EIAEDs

Patient	Dose (mg)	CSS, max (ng/mL)	AUCSS (ng × h/mL)	Dose Normalized AUCSS (ng × h/mL)
1	300	586	11 597	39
7	300	675	12 450	41
9	300	728	14 342	48
19	400	1200	23 347	58
20	400	1190	22 275	56
27	300	722	5278	18
32	500	799	17 523	35
39	500	1490	28 268	57

Fig. 1.

Individual steady-state (_ss) AUC values against dose in 8 patients on the phase I trial. There is a trend for the steady-state exposure to vandetanib to increase with dose.

Fig. 2.

Dose normalized individual steady-state (_ss) AUC values against dose in historical controls treated with vandetanib, not on EIAEDs, compared with 8 patients on the phase I study, on EIAEDs. There is no significant difference between these 2 groups, suggesting that there is no decrease in exposure to vandetanib in patients on EIAEDs.

Phase II

Patient characteristics

Between January 2006 and August 2009, 64 patients were accrued to the phase II portion of the study (32 GBM and 32 AG). Seven patients in the AG arm had an oligodendroglial component to their tumor. Of 5 tested, 2 patients had 1p19q deletion. Fourteen patients in the AG arm and 5 patients in the GBM arm had received VEGF-directed therapy prior to enrollment in this trial. All patients had received prior radiotherapy, and the majority of patients were treated for their first or second recurrence (range, 1–8). All patients were treated at the National Institutes of Health Clinical Center. Detailed patient characteristics are described in Table 3. One patient in the AG arm was registered but excluded from analysis because he withdrew consent after one dose of drug.

Table 3.

Patient characteristics for the phase II study

Characteristic	GBM, N = 32	AG, N = 32
Sex
Male	20 (63%)	24 (75%)
Female	12 (37%)	8 (25%)
Age, y
Median	48	41
Range	21–70	18–69
KPS
Median	90	90
Range	60–100	60–100
Prior therapy
Radiation	32 (100%)	32 (100%)
Prior chemotherapy regimens	32 (100%)	32 (100%)
Median	1	2
Range	1–8	1–6
First recurrence	18 (56%)	10 (31%)
Second recurrence	7 (22%)	4 (13%)
3+ Recurrences	7 (22%)	18 (56%)
Prior VEGF-directed therapy	5 (16%)	13 (41%)
Steroids
Yes	19 (60%)	16 (50%)
No	13 (40%)	16 (50%)
Anti-epileptics
Yes	21 (66%)	24 (75%)
No	11 (34%)	8 (25%)

Toxicity

Vandetanib was moderately toxic in this cohort of patients. The most common grade 1–2 treatment-related toxicities were rash (63%), diarrhea (53%), prolonged QTc (48%), hypertension (33%), and elevated aspartate aminotransferase (31%). Toxicities related to grades 3–4 included seizure (16%), rash (6%), and lymphopenia (6%). Other isolated severe adverse events included hypertension, thromboembolic events, erythema multiforme, CNS hemorrhage, Fournier's gangrene, gastrointestinal perforation, and cerebrovascular ischemia. Twelve patients discontinued vandetanib due to toxicity, and 5 patients required dose reductions. Data on treatment-related grades 3 and 4 toxicities are in Table 4.

Table 4.

Grade 3 and 4 toxicities related to treatment with vandetanib in 64 patients

Toxicity	Grade 3	Grade 4	Grade 5	Total
Seizure	9	1	0	10
Lymphopenia	3	1	0	4
Rash	4	0	0	4
Prolonged QTc	2	0	0	2
Hypertension	2	0	0	2
Fatigue	0	1	0	1
Thrombocytopenia	1	0	0	1
Neutropenia	1	0	0	1
Hemorrhage, CNS	0	0	1	1
Anorexia	1	0	0	1
Skin breakdown–decubitus ulcer	0	1	0	1
Thrombosis/embolism	1	1	0	1
Rash–erythema multiforme	0	1	0	1
Gastrointestinal perforation, colon	1	0	0	1

Three patients suffered symptomatic CNS hemorrhages while on study. One patient on anticoagulation had progression of a parenchymal hemorrhage 8 weeks into treatment. A second patient, not on anticoagulation, suffered a 3 × 4-cm parenchymal hemorrhage 8 weeks into the study. Finally, a third patient died following a large parenchymal hemorrhage into an area of previous petechial hemorrhage 5 weeks into the study. There were no other deaths related to vandetanib.

Efficacy

Thirty-two patients were included in the PFS and OS analyses in the GBM arm. The median number of cycles of vandetanib was 2 (range, 1–73). Twenty-seven patients (84%) stopped vandetanib due to tumor progression, 3 (9%) due to toxicity, and 1 due to intercurrent illness unrelated to vandetanib. One patient is still on trial. One patient was censored for progression because of stopping therapy for toxicity without evidence of progression. PFS6 was 6.5% (95% confidence interval [CI]: 1.7%–24.7%). Median PFS was 1.3 months (95% CI: 0.9–1.9 mo). Two patients were progression free at 6 months, but both of these patients had been enrolled within 12 weeks of radiation. One patient who is still alive and on study was censored for survival at the date of last evaluation. All others had died by the time of analysis. Median OS was 6.3 months (95% CI: 3.9–8.5 mo). Results are summarized in Table 5.

Table 5.

Results of the phase 2 study

	GBM	AG
PFS6	6.5% (95% CI: 1.7%–24.7%)	7.0% (95% CI: 1.8%–26.4%)
Median PFS	1.3 mo (95% CI: 0.9–1.9 mo)	1.8 mo (95% CI: 0.9–2.7 mo)
OS 6 mo	56.3% (95% CI: 41.4%–76.4%)	56.3% (95% CI: 41.4%–76.4%)
Median OS	6.3 mo (95% CI: 3.8–8.5 mo)	7.6 mo (95% CI: 4.5–12.5 mo)

Twenty-seven patients were evaluable for radiographic response in the GBM arm. Four patients (15%; 95% CI: 4.2%–33.7%) had an objective radiographic response to therapy. The response rate was 12.5% (95% CI: 3.5%–29.0%) in the intention-to-treat cohort of 32 patients. Two of those patients had been enrolled within 12 weeks of chemoradiation, and resolution of pseudoprogression cannot be ruled out. One patient had a complete response maintained for over 5 years and remains on study. (Fig. 3) This patient was treated within 90 days of completing radiation, and although resection prior to study entry confirmed residual viable tumor with a MIB-1 index of 10%, results in this patient should be interpreted with caution.

Fig. 3.

(A) Baseline study, patient 6. (B) Maintained complete response after >5 y of therapy with vandetanib for recurrent GBM. (C) Baseline study, patient 77. (D) Partial response after 16 wk of therapy with vandetanib for recurrent anaplastic astrocytoma.

Thirty-two patients were included in the PFS and OS analyses in the AG arm. The median number of cycles of vandetanib was 2 (range, 1–44). Twenty-three (72%) stopped vandetanib due to tumor progression, 5 (16%) due to toxicity, and 1 due to intercurrent illness unrelated to vandetanib. Two (6%) withdrew for reasons other than toxicity, and 1 is still on trial. Three patients were censored for progression. Two patients refused further treatment and were subsequently lost to follow-up—they both had stable disease documented by MRI at the time they came off study and were censored at the date of their last evaluation. One stopped therapy for toxicity related to study drug and was censored on the date he started a new therapy without interval progression. PFS6 was 7.0% (95% CI: 1.8%–26.4%). Median PFS was 1.8 months (95% CI: 0.9–2.7 mo). Three patients were alive at the time of analysis and were censored for survival at the date of their last evaluation. Median OS was 7.6 months (95% CI: 4.5–12.5 mo).

Thirty patients were evaluable for radiographic response in the AG arm. Two patients (7%; 95% CI: 0.8%–22.1%) had an objective response to therapy. Response rate was 6.3% (95% CI: 0.8%–20.8%) in the intention-to-treat cohort. One patient had no measurable enhancing disease when he started the trial and has remained stable for over 3 years on study (Fig. 3). Five additional patients achieved stable disease for at least 2 months.

As a secondary analysis, we explored whether several common toxicities associated with vandetanib (rash, hypertension, and seizure) could be used as biomarkers to predict response to treatment. None of these toxicities was significantly correlated with either OS or PFS6 in the GBM or AG arms, or with radiographic response in the entire cohort.

Discussion

Vandetanib is a logical candidate to explore in the investigation of novel antiglioma therapies. Combination VEGF and EGFR pathway blockade is being explored in other solid tumor malignancies, such as renal and lung carcinomas.^26–29 Relative to malignant gliomas, a phase II trial was recently reported where 25 patients with recurrent GBM were treated with bevacizumab and erlotinib. Investigators observed a PFS6 of 24% and a response rate approaching 50%.³⁰ In the current trial, combination VEGF and EGFR pathway blockade was accomplished using an orally available, single-agent multitargeted tyrosine kinase inhibitor, although with less robust results. There are several potential reasons for the lower response rate seen in our trial. Bevacizumab may provide more complete angiogenesis inhibition through its binding of the VEGF ligand, in contrast to vandetanib, which selectively inhibits VEGFR-2 activity. Indeed, in trials,^31,32 other single-agent VEGFR small molecule inhibitors have proven less active in GBM than has bevacizumab. Additionally, 4 GBM patients and 13 anaplastic astrocytoma patients enrolled in this study had progressed on bevacizumab prior to treatment with vandetanib, and 1 GBM patient had received prior sorafenib (VEGFR, platelet-derived growth factor receptor, Raf). None of these patients had a response to therapy and therefore may already have acquired resistance to VEGF inhibitor therapy prior to coming onto trial. Their OS was significantly shorter, in both the GBM cohort (median OS 4.9 vs 10.4 mo, log-rank P = .0197) and the AG cohort (median OS 4.0 vs 12.4 mo, log-rank P = .0016), supporting this hypothesis. Confounders may be that these patients were also older and more heavily pretreated. Thus, vandetanib may have proven to be more active in a patient population naive to anti-VEGF therapy. In an attempt to test this hypothesis, a phase II study of low-dose vandetanib (100 mg daily) in combination with radiation and temozolomide in patients with newly diagnosed GBM is ongoing, although interim results do not show a significant improvement in PFS or OS compared with standard chemoradiation.³³

While durable disease control was not observed in this study cohort, a number of responses to treatment were seen, with long-term benefit for 1 patient who had a complete radiographic response for over 5 years, and another who has had no measurable disease for over 3 years. Interestingly, all patients with objective responses to vandetanib also had rash as a complication of therapy. Rash has been described as a predictor of response to erlotinib and gefitinib treatment for non-small-cell lung carcinoma patients.^34,35 While the same correlation has not been observed to a significant degree in glioma patients receiving EGFR inhibitor monotherapy, rash may be a relevant biomarker for patient outcomes with vandetanib and warrants further study.

An unexpected treatment-related adverse event was an apparent increased frequency of seizures, usually within the first cycle of therapy in patients initiating treatment with vandetanib. It is difficult to quantify the increased risk for seizures with vandetanib because patients with primary brain tumors are at inherent increased risk for seizures, especially in the setting of disease progression. Nevertheless, there were several cases of patients having a generalized seizure as their first-ever seizure during the first cycle of vandetanib and a number of other patients experiencing a significantly increased frequency of seizures after starting vandetanib and then returning to baseline after terminating vandetanib treatment despite disease progression. Thus, it appears that vandetanib does increase the likelihood of seizures, a toxicity that may be specific to vulnerable brain tumor patients because increased seizures have not been described in other clinical trials of vandetanib. Otherwise, the toxicity profile was similar to that described in larger phase III studies of vandetanib,¹⁸ with the exception of lymphopenia, which may be a treatment-emergent toxicity and related to the frequent use of long-term steroids in this patient population.

PK data in a subset of patients on the phase I portion of the trial showed no significant difference in dose-normalized AUC compared with historical patients treated with a 300-mg dose, suggesting that EIAEDs may not have a significant effect on vandetanib exposure. A recent study did show a significant reduction (40%) in the AUC after a single dose of vandetanib in patients who were coadministered rifampicin, a potent CPY3A4 inducer, and a significant (9%) increase in the AUC with coadministration of itraconazole, a potent CPY3A4 inhibitor.²¹ Patients on the phase I trial were treated with a variety of EIAEDs with varying degrees of CYP3A4 induction. Patients on carbamazepine were overrepresented in our PK study, so results may not apply to patients on more potent inhibitors such as phenytoin.

In conclusion, vandetanib was a moderately tolerated drug but did not have significant activity as a single agent in patients with recurrent malignant glioma.

Funding

This work was supported by the National Cancer Institute (NCI) Intramural Research Program.